Objective:To explore the application value of machine learning algorithms in constructing a predictive model for cardiovascular toxicity in breast cancer patients receiving anthracycline-based chemotherapy.Methods:This study was a retrospective cohort study. The female patients with breast cancer who received anthracyclines in the Affiliated Cancer Hospital of Xinjiang Medical University from January 2020 to December 2023 were enrolled. The endpoint event was abnormal electrocardiogram (ECG). According to whether the patients had ECG abnormalities during chemotherapy, they were divided into the ECG abnormal group and the ECG normal group. The dataset was divided into the training set and the test set at a ratio of 8∶2, and logistic regression, random forest, extreme gradient boosting (XGBoost), support vector machine (SVM) and multilayer perceptron (MLP) were used to construct a risk prediction model for cardiovascular toxicity in breast cancer patients, and the receiver operating characteristic curve, calibration curve and clinical decision curve were used to evaluate the model.Results:A total of 731 female patients with breast cancer, aged (51.6±9.4) years, were enrolled. The follow-up time was (130.3±37.1) days. There were 333 cases in the ECG abnormal group and 398 cases in the ECG normal group. Seven factors influencing cardiovascular toxicity were identified, including age, menstrual history, diabetes, combination therapy with trastuzumab, combination therapy with dexrazoxane, creatine kinase isoenzymes, and α-hydroxybutyrate dehydrogenase. In the training set, the area under the curve ( AUC) for the logistic regression, random forest, XGBoost, SVM, and MLP models was 0.712, 0.863, 0.774, 0.813, and 0.733, respectively. In the test set, the AUC was 0.671, 0.778, 0.746, 0.771, and 0.705, respectively. Calibration curves and clinical decision curves showed that the random forest model performed the best. Conclusion:Models constructed with machine learning algorithms show promise in predicting cardiovascular toxicity in breast cancer patients receiving anthracycline-based chemotherapy, with the random forest prediction model performing the best.

Objective:To explore the application value of machine learning algorithms in constructing a predictive model for cardiovascular toxicity in breast cancer patients receiving anthracycline-based chemotherapy.Methods:This study was a retrospective cohort study. The female patients with breast cancer who received anthracyclines in the Affiliated Cancer Hospital of Xinjiang Medical University from January 2020 to December 2023 were enrolled. The endpoint event was abnormal electrocardiogram (ECG). According to whether the patients had ECG abnormalities during chemotherapy, they were divided into the ECG abnormal group and the ECG normal group. The dataset was divided into the training set and the test set at a ratio of 8∶2, and logistic regression, random forest, extreme gradient boosting (XGBoost), support vector machine (SVM) and multilayer perceptron (MLP) were used to construct a risk prediction model for cardiovascular toxicity in breast cancer patients, and the receiver operating characteristic curve, calibration curve and clinical decision curve were used to evaluate the model.Results:A total of 731 female patients with breast cancer, aged (51.6±9.4) years, were enrolled. The follow-up time was (130.3±37.1) days. There were 333 cases in the ECG abnormal group and 398 cases in the ECG normal group. Seven factors influencing cardiovascular toxicity were identified, including age, menstrual history, diabetes, combination therapy with trastuzumab, combination therapy with dexrazoxane, creatine kinase isoenzymes, and α-hydroxybutyrate dehydrogenase. In the training set, the area under the curve ( AUC) for the logistic regression, random forest, XGBoost, SVM, and MLP models was 0.712, 0.863, 0.774, 0.813, and 0.733, respectively. In the test set, the AUC was 0.671, 0.778, 0.746, 0.771, and 0.705, respectively. Calibration curves and clinical decision curves showed that the random forest model performed the best. Conclusion:Models constructed with machine learning algorithms show promise in predicting cardiovascular toxicity in breast cancer patients receiving anthracycline-based chemotherapy, with the random forest prediction model performing the best.